M

Ct

LMa

Bb

c

a

ARRAA

KCMNXX

I

mdo2oi“t

3Jrfc

(

0h

Journal of Plant Physiology 170 (2013) 934– 942

Contents lists available at SciVerse ScienceDirect

Journal of Plant Physiology

journa l h o me page: www.elsev ier .com/ locate / jp lph

olecular biology

haracterization of Citrus sinensis transcription factors closely associated withhe non-host response to Xanthomonas campestris pv. vesicatoria

ucas D. Daurelioa, María S. Romeroa, Silvana Petrocelli a, Paz Merelob, Adriana A. Cortadic,anuel Talónb, Francisco R. Tadeob,∗, Elena G. Orellanoa,∗∗

Área de Biología Molecular, Instituto de Biología Molecular y Celular de Rosario (IBR), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Facultad de Cienciasioquímicas y Farmacéuticas, Universidad Nacional de Rosario (UNR), Suipacha 531 (S2002LRK), Rosario, Santa Fe, ArgentinaCentre de Genómica, Institut Valencià d’Investigacions Agràries (IVIA), Apt. Oficial, 46113 Montcada, València, SpainÁrea de Biología Vegetal, Facultad de Ciencias Bioquímicas y Farmacéuticas, Universidad Nacional de Rosario (UNR), Suipacha 531 (S2002LRK), Rosario, Santa Fe, Argentina

r t i c l e i n f o

rticle history:eceived 11 October 2012eceived in revised form 29 January 2013ccepted 29 January 2013vailable online 26 February 2013

eywords:itrus sinensisicroarray transcriptome analysis

a b s t r a c t

Plants, when exposed to certain pathogens, may display a form of genotype-independent resistance,known as non-host response. In this study, the response of Citrus sinensis (sweet orange) leaves to Xan-thomonas campestris pv. vesicatoria (Xcv), a pepper and tomato pathogenic bacterium, was analyzedthrough biochemical assays and cDNA microarray hybridization and compared with Asiatic citrus cankerinfection caused by Xanthomonas citri subsp. citri. Citrus leaves exposed to the non-host bacterium Xcvshowed hypersensitive response (HR) symptoms (cell death), a defense mechanism common in plants butpoorly understood in citrus. The HR response was accompanied by differentially expressed genes that areassociated with biotic stress and cell death. Moreover, 58 transcription factors (TFs) were differentially

on-host responseanthomonas campestris pv. vesicatoriaanthomonas citri subsp. citri

regulated by Xcv in citrus leaves, including 26 TFs from the stress-associated families AP2-EREBP, bZip,Myb and WRKY. Remarkably, in silico analysis of the distribution of expressed sequence tags revealedthat 10 of the 58 TFs, belonging to C2C2-GATA, C2H2, CCAAT, HSF, NAC and WRKY gene families, werespecifically over-represented in citrus stress cDNA libraries. This study identified candidate TF genes forthe regulation of key steps during the citrus non-host HR. Furthermore, these TFs might be useful infuture strategies of molecular breeding for citrus disease resistance.

ntroduction

Plants are constantly exposed to a variety of pathogenicicroorganisms. The results of plant–pathogen interactions

epend on the combination of interacting partners, the plant devel-pmental stage and environmental signals (Ascencio-Ibanez et al.,008). A plant species is a “host” for a particular pathogen if at leastne cultivar is susceptible to it. This dynamic, which leads to illness,

s called “compatible interaction”. Phytopathogens exhibit narrowhost” specificity and are unable to infect “non-host” plants dueo non-host defense, a response considered “incompatible”. Plant

Abbreviations: CFU, colony forming units; Ctr, negative control; DAB,-3diaminobenzidine; EST, expressed sequence tag; HR, hypersensitive response;A, jasmonic acid; LCD, localized cell death; pi, post infiltration; PR, pathogenesis-elated protein; ROS, reactive oxygen species; SA, salicylic acid; TF, transcriptionactor; Xcv, Xanthomonas campestris pv. vesicatoria; Xcc, Xanthomonas citri subsp.itri.∗ Corresponding author. Tel.: +34 963424129; fax: +34 963424000.

∗∗ Corresponding author. Tel.: +54 3414350661x106; fax: +54 3414390465.E-mail addresses: tadeo fra@gva.es (F.R. Tadeo), orellano@ibr.gov.ar

E.G. Orellano).

176-1617/$ – see front matter © 2013 Elsevier GmbH. All rights reserved.ttp://dx.doi.org/10.1016/j.jplph.2013.01.011

© 2013 Elsevier GmbH. All rights reserved.

non-host resistance exhibits constitutive and inducible featuresthat are not well understood at the molecular level (Nürnbergerand Lipka, 2005). Non-host defense is recognized as a hypersensi-tive response (HR) if it presents visible symptoms. These symptomsresult in localized cell death (LCD), with necrotic lesions thatconfine the pathogen at the infection site (Oh et al., 2006; Muret al., 2008; Daurelio et al., 2011). The initial HR events followingpathogen recognition include structural modifications, cytoplas-mic membrane disruption, production of reactive oxygen species(ROS), synthesis of pathogenesis-related proteins (PRs) and trans-criptional reprogramming (Eulgem, 2005; Oh et al., 2006; Daurelioet al., 2009b). The type of plant response depends on quantitativetranscriptome modifications (Tao et al., 2003). Transcription fac-tors (TFs) therefore play a fundamental role, and several of themhave been shown to regulate the expression of defense-relatedgenes (Singh et al., 2002; Eulgem, 2005). Accordingly, microarraytechnology has helped unravel key processes in plant responsesto pathogens (Eulgem, 2005). The comprehension of transcriptio-

nal modifications involved in plant non-host resistance representsboth a great scientific achievement and a fundamental step towardidentifying genes for plant disease control breeding programs(Singh et al., 2002; Daurelio et al., 2011).

lant P

wLmAisdfalrrel(ef2frs2awGryXsHrsa

ttpsoabsTdm

M

P

vAwh

(B(fctau

L.D. Daurelio et al. / Journal of P

Although plant defensive mechanisms involve general path-ays, they differ based on important traits between genera (van

oon et al., 2006). Moreover, while numerous studies have analyzedolecular defense mechanisms in herbaceous model plants such as

rabidopsis, tobacco, soybean and rice, little is known about themn tree crops of economic importance (Gomi et al., 2003). Citruspecies are the most economically important fruit crop in the worldue to their high productivity and the nutrient content of theirruits, yet they suffer devastating losses caused by pathogens (Talonnd Gmitter, 2008). The transcriptional response of lemon (C. limon)eaves against non-pathogenic strains of Alternaria alternata (a cit-us pathogenic fungus) using subtractive hybridization detected aeduced number of pathways involved in this interaction (Gomit al., 2003). Recent advances in functional genomics and molecu-ar biology, including the analysis of large expressed sequence tagEST) collections in public databases (Forment et al., 2005; Campost al., 2007; Guidetti-Gonzalez et al., 2007) and the design of dif-erent microarrays platforms (Fujii et al., 2007; Cernadas et al.,008; Martinez-Godoy et al., 2008), have provided valuable toolsor citrus–pathogen interaction research. Consequently, the cit-us transcriptome was recently analyzed using microarrays duringome compatible interactions (Mozoruk et al., 2006; Gandia et al.,007; Cernadas et al., 2008; Kim et al., 2009). In addition, in siliconalyses of PR- and HR-related ESTs have been carried out, but theseere not confirmed experimentally (Campos et al., 2007; Guidetti-onzalez et al., 2007). To our knowledge, neither citrus non-host

esponse characterization nor citrus non-host transcriptome anal-ses have been carried out to date. Although the citrus response toanthomonas fuscans pv. aurantifolli was defined as non-host, theymptoms manifested as an arrested canker instead of a typicalR (Cernadas et al., 2008). While the Key lime response to the cit-

us canker producing bacteria Xanthomonas citri subsp. citri (Xcc)train Aw showed HR symptoms, the plant transcriptome was notnalyzed during this host interaction (Rybak et al., 2009).

We studied the interaction between sweet orange leaves andhe Gram-negative bacterium Xanthomonas campestris pv. vesica-oria (Xcv), the agent responsible for spot disease in tomatoes andeppers (Thieme et al., 2005). The interaction was previously con-idered to be an HR (Daurelio et al., 2009a); this was confirmed inur study. In addition, we performed a microarray transcriptomenalysis comparing the citrus–Xcv interaction against the compati-le citrus–Xcc interaction to detect genes with modified expressionpecific to the citrus non-host response. Biotic stress, cell death andF genes differentially expressed in the citrus response to Xcv wereescribed. This study represents an initial effort to elucidate theechanisms underlying the citrus non-host response to pathogens.

aterials and methods

lant material, bacterial strains and inoculation procedure

Citrus sinensis L. Osbeck cv. Valencia Late plants generously pro-ided by Catalina Anderson and Gastón Alanis (INTA Concordia,rgentina) were grown in a greenhouse at 25/18 ◦C (day/night)ith a 14 h photoperiod (150 �E m−2 s−1) and controlled relativeumidity. Young, fully expanded leaves were used in all assays.

X. citri subsp. citri (Hasse) and X. campestris pv. vesicatoriaDoidge) strains were kindly provided by Blanca I. Canteros (INTAella Vista, Argentina). Strains were grown at 28 ◦C in SB mediumDaurelio et al., 2011) supplemented with 25 �g/mL ampicillin

or Xcc. Bacterial overnight cultures were diluted to the indicatedolony forming units (CFU)/mL in 10 mM MgCl2 and pressure infil-rated into the abaxial side of the leaves using a syringe without

needle (Daurelio et al., 2009b). A solution of 10 mM MgCl2 wassed as a negative control (Ctr).

hysiology 170 (2013) 934– 942 935

Infectivity study, histological analysis, ion leakage and growthcurves

In the infectivity assays, phenotypes were discerned visually.To perform histological analysis, an inoculated region of approxi-mately 1 cm2 was excised at 48 h post infiltration (pi) and fixed inFAA solution (10% formaldehyde, 5% glacial acetic acid, 50% ethanol)for 24 h. After fixation, samples were dehydrated and embeddedin paraffin (D’Ambrogio de Argüeso, 1986). Cross-sections wereplaced on glass slides, and paraffin was removed with xylene. Sam-ples were stained with hematoxylin-eosin (Johansen, 1940) forobservation by light microscopy.

Inoculated leaves were cut 2 hpi. Petioles were submerged in1 mg/mL 3-3 diaminobenzidine (DAB), pH 3.8, for 18 h in darknessto detect hydrogen peroxide (H2O2) production (Daurelio et al.,2009b). Leaves were subsequently bleached and stored in 96%ethanol at room temperature.

Ion leakage measurements were performed to detect cytoplas-mic membrane disruption. Leaf discs with a diameter of 1.3 cmwere taken at 0, 4, 8 and 24 hpi. Leaf discs were rinsed for 10 minin 4 mL distilled water and incubated for 3 h in 1 mL fresh distilledwater. Electrical conductivity was referred to as 100% ion content inboiled leaf discs (120 ◦C, 20 min) (Daurelio et al., 2009b). As a mem-brane disruption control, 1 mM salicylic acid (SA) was included.Assays were run in triplicate in independent leaves. A three-factor(treatment, time pi and leaf) mixed model ANOVA and a Tukey’smultiple comparison test were used for statistical analyses. Resid-ual analyses and validation with logarithmic data transformationwere performed.

In planta growth assays were performed by grinding 1 cm2 infil-trated leaf discs in 100 �L distilled water. These solutions wereserially diluted and plated onto SB agar. Colonies were counted after48 h of incubation at 28 ◦C and expressed as log CFU/cm2 leaf tissue.

RNA isolation, sample labeling and microarray hybridization

Regions inoculated with Xcv, Xcc (107 CFU/mL) or Ctr were col-lected at 8 hpi. Tissue was harvested in liquid nitrogen and stored at−80 ◦C until RNA isolation. Total RNA was extracted using TRIzol®

Reagent (Invitrogen) according to the manufacturer’s protocol. RNAquality was tested using the OD260/OD280 ratio and agarose gelelectrophoresis (Sambrook et al., 1989). First-strand cDNA wasgenerated from 1 �g total RNA. Double-stranded cDNA synthe-sis, in vitro transcription and amplification, labeling of amplifiedRNA, microarray hybridization and slide washes were performed asdescribed by Cercós et al. (2006). RNA was extracted from three bio-logical replicates, independently processed, labeled and hybridizedto different microarrays. Individual RNA samples were used tosynthesize Cy5-labeled antisense cRNA. These samples were co-hybridized with Cy3-labeled antisense cRNA that was synthesizedfrom a reference sample containing a mixture of equal amounts ofamplified RNA from all experimental samples (Agusti et al., 2008).Samples were labeled with Cy5 and the reference with Cy3 to avoiddye artifacts. A citrus cDNA microarray containing 21,081 putativeunigenes was utilized (Martinez-Godoy et al., 2008).

Data acquisition and analysis

Hybridized arrays were scanned using a Scanarray Gx scanner(PerkinElmer) with Scanarray Express software following the man-ufacturer’s instructions. An appropriate photomultiplier gain ratiofor the two channels and a percentage of 1% of the saturated spots

were obtained. GenePix 4.1 (Axon Instruments) was used to trans-form intensity measurements into numeric data. Data from spotsflagged as “not found” or “bad” during scanning and spots with asignal to background ratio lower than 2 were discarded. Because

9 lant P

tstsi(eserfXVFoIg

st1≥dat(

aao(uTwbTcr

F(l

36 L.D. Daurelio et al. / Journal of P

he comparisons performed revealed a low percentage of probeshowing significant differences in expression, robust normaliza-ion was carried out using the Lowess method. Probes showingignificant differential gene expression between samples weredentified using the Linear Models in Microarrays (LIMMA) packageSmith, 2004) from the Bioconductor software library (Gentlemant al., 2004). Gene expression differences were considered to beignificant when p-values (considering correction for false discov-ry rate) were lower than 0.01. The M value (log2 of expressionatio between treatments) cutoff was ±1 and was M = log2[Xcv/Ctr]or Xcv/Ctr, M = log2[Xcc/Ctr] for Xcc/Ctr and M = log2[Xcv/Xcc] forcv/Xcc comparisons. Venn diagrams were constructed using theenny tool (http://bioinfogp.cnb.csic.es/tools/venny/index.html).unctional classification of the selected EST subgroups was carriedut over their Arabidopsis thaliana orthologs through MIPS (Munichnformation Center for Protein Sequences, http://www.mips.sf.de) and MapMan software (Thimm et al., 2004).

Full-length sequences of selected unigenes were made byequential assembly of homolog ESTs. ESTs were retrieved usinghe Basic Local Alignment Search Tool BLASTN (Altschul et al.,997), considering a query coverage of ≥10% and a max identity of95%, from the National Center for Biotechnology Information ESTatabase (http://www.ncbi.nlm.nih.gov/dbEST). Sequences weressembled using the CAP3 program (Huang and Madan, 1999). Pro-ein sequences and domains were determined using ORF finderStothard, 2000) and Pfam (Finn et al., 2006) software, respectively.

Genes that were preferentially expressed in libraries associ-ted with stress were identified using statistically based in siliconalysis. A total of 514,845 citrus ESTs were categorized as stressr non-stress based on conditions used for library constructionsee Supplementary Table S1). Paired comparisons were conductedsing the Chi-squared test with IDEG6 (Romualdi et al., 2003).he numbers of stress/non-stress ESTs corresponding to each geneere used as the input parameters, in addition to the total num-

er of ESTs grouped into each category (Supplementary Table S1).he differences in gene expression under stress conditions wereonsidered statistically significant according to the Bonferroni cor-ection.

ig. 1. Characterization of the citrus response to Xcv. (A) Inoculated leaves showing phenoB) Histological analysis by hematoxylin-eosin staining showing leaf structure infiltratedeaves. Values are averages of CFU/cm2 recovered. Bars represent SD (n = 3). Inoculated ba

hysiology 170 (2013) 934– 942

Phylogenetic trees (5000 bootstrap) based on the neighbor-joining method (Saitou and Nei, 1987) were generated usingMEGA4 (Tamura et al., 2007). The representative domains forcitrus and Arabidopsis alleles used as input sequences were pre-viously detected by in-batch Pfam analysis and extracted using aDaurelio–Kurth developed Perl script (www.perl.org).

Real-time RT-PCR

Gene expression was confirmed by real-time RT-PCR analysis.Primers were designed using Primer3 v.0.4.0 software (Rozen andSkaletsky, 2000). Analyzed ESTs, primers and product lengths areindicated in Supplementary Fig. S1. One microgram of total RNAfrom the same samples used in the microarray experiments wasused for cDNA synthesis with the M-MuLV Retro Transcriptaseenzyme (Promega, USA) and d(T)22 oligonucleotide, following themanufacturer’s instructions. PCR products using genomic DNA orcDNA templates for the actin housekeeping gene were sized differ-ently, allowing for the detection of genomic DNA contamination.PCR reactions without the reverse transcription step did not yieldproducts. Real-time PCR, specificity analysis using melting curvesand data normalization to actin were performed as described inDaurelio et al. (2011).

Results

Characterization of the orange leaf–Xcv interaction

The response of Citrus sinensis leaves to Xcv treatment wascompared with the response to Xcc treatment as the disease ref-erence. Orange leaves showed necrosis in infectivity assays using106, 107 and 108 CFU/mL Xcv inoculums between 48 and 72 hpi.Some chlorosis was observed when leaves were infiltrated withXcv 105 CFU/mL, Xcc or Ctr (Fig. 1A). The initial symptoms of water

soaking were observed in the abaxial region of leaves inoculatedwith Xcc (Fig. 1A).

The non-host response phenotype was confirmed at the cellularlevel by histological staining at 48 hpi. In comparison with Ctr, no

typic responses on the abaxial (left) and adaxial (right) sides are observed at 72 hpi. with Ctr, Xcc and Xcv at 48 hpi (Bar = 50 �m). (C) Growth curves of Xcv in orangecteria dilutions are shown as CFU/mL.

L.D. Daurelio et al. / Journal of Plant Physiology 170 (2013) 934– 942 937

F owinge d. Vaa

mic

epliu

eaaio

Ir

i

FoD

ig. 2. Characterization of the citrus response to Xcv. (A) DAB staining at 2 hpi shlectrolytic conductivity for the treatments and non-infiltrated regions (NI) exhibites CFU/mL.

icroscopic changes were observed in the Xcc interaction. Xcv-noculated leaves presented great areas of mesophyll cell lysis andomplete structure disorganization (Fig. 1B).

During the study of Xcv behavior in planta, the number of recov-red bacteria remained constant until 2 dpi. At 6 dpi, the bacterialopulation had diminished by two orders of magnitude for the

argest initial inoculum and by one order for the smallest initialnoculum. In both cases, bacterial populations decreased slowlyntil 10 dpi (Fig. 1C).

Dark staining due to DAB oxidation by peroxidases in the pres-nce of H2O2 was observed in Xcv inoculated leaves at 4 hpi. Xccnd Ctr treatments showed no such reaction (Fig. 2A). In ion leakagessays, no differences were observed between Ctr, Xcc and non-nfiltrated regions. However, significant ion leakage increases werebserved for Xcv at 8 and 24 hpi (Fig. 2B, p < 0.001).

dentification of differentially expressed genes during the citrus

esponse to Xcv

Gene expression profiling was carried out on bacterial-nfiltrated Citrus sinensis leaf tissue at 8 hpi using cDNA microarrays.

ig. 3. Venn diagrams showing the number of orange ESTs differentially regulated at 8 hpif ESTs are shown in boxes. Grey zones represent ESTs differentially expressed between Xcata are based on microarray analyses.

H2O2 localization in leaf tissue treated with Ctr, Xcc and Xcv. (B) Percentages oflues are averages. Bars represent SD (n = 3). Inoculated bacteria dilutions are shown

Of the 21,081 putative unigenes in the citrus cDNA microarray,2439 (11.5%) ESTs showed altered expression in both bacterialtreatments, 1022 (42%) ESTs were induced, and 1417 (58%) wererepressed (Fig. 3). Most of these genes were differentially reg-ulated in the Xcv treatment compared with the Ctr (930 ESTsinduced, 1309 ESTs repressed) or in the Xcv treatment comparedwith the Xcc treatment (732 ESTs induced, 813 ESTs repressed).Fewer differentially regulated ESTs were observed in the Xcctreatment compared to Ctr (64 ESTs induced, 71 ESTs repressed,Fig. 3).

Because the aim of this study was to detect genes specifi-cally involved in the citrus non-host response, we evaluated thosegenes that were differentially expressed in Xcv when comparedto Xcc and Ctr treatments (Fig. 3). Considering both subgroups,654 probes were induced and 733 were repressed. Of these, 22induced probes and 5 repressed probes were also differentiallyexpressed in Xcc treatment compared to Ctr (Fig. 3). The respec-

tive Arabidopsis orthologs were assigned and functionally grouped(data not shown). Genes related to biotic stress and cell deathand TF genes were analyzed with regard to the non-host responseprocess.

in leaf tissue during response to Xcv compared to Xcc treatment. The total numbersv/Xcc and Xcv/Ctr, or between Xcv/Xcc, Xcv/Ctr and Xcc/Ctr (white boxed numbers).

938 L.D. Daurelio et al. / Journal of Plant P

FXs

B

wTl

TCc

ig. 4. Citrus ESTs that are differentially regulated during the non-host response tocv related to biotic stress. The log2 of expression ratio between treatments (M) arehown. Data are averages of three repetitions. Bars represent SE.

iotic stress-related genes

Nineteen probes corresponding to 18 biotic stress-related genesere found to be differentially expressed (Fig. 4, Supplementary

able S2). Twelve of these genes showed homology to PRs as fol-ows: two PR-3 and two PR-4 endochitinases that protect plants

able 1itrus ESTs that are differentially regulated during the non-host response to Xcv relateontrast cutoff values of ±1 are shown in bold. Data are averages of three repetitions.

ID Putative gene identification [Species](Accession No.)

E-value

C02006G06 Cyclic nucleotide-gated ion channel, putative[Ricinus communis] (XP 002532894)

0E+00

C02011G07 Cyclic nucleotide-gated ion channel, putative[Ricinus communis] (XP 002525760)

0E+00

C31204D05 Red chlorophyll catabolite reductase ACD2,chloroplast precursor, putative [Ricinuscommunis] (XP 002523576)

4E−104

C31401A10 Unnamed protein product, putative CAD1 [Vitisvinifera] (CBI18046)

0E+00

KN0AAI1AA02 Predicted protein, putative cysteine proteaseinhibitor [Populus trichocarpa] (XP 002336103)

2E−16

hysiology 170 (2013) 934– 942

against fungi, nematodes and insects; three thaumatin-like PR-5 proteins with activity against oomycetes; one PR-6 proteinaseinhibitor targeting nematodes and insects; two PR-10 enzymeswith ribonuclease activity against viruses; one PR-13 or thionin-like protein with broad antibacterial and antifungal activity; andone PR-15 or germin-like protein (van Loon et al., 2006). Theother six genes had the following homologies: two Mlo receptors(Mla powdery-mildew-resistant), a Ve-like (Verticillium wilt dis-ease) leucine-rich repeat receptor protein, a lipase containing adomain similar to PAD4 (Phytoalexin Deficient 4), the NADPH oxi-dase RBohD and the inducible protein Hin-1 (hairpin type 1). Allgenes were induced in the Xcv treatment, except the germin-likeprotein (PR-15) and the homolog of the Ve-like receptor (Fig. 4).

PR-15 and Mlo6 were significantly under- and over-represented, respectively, under stress conditions. These resultswere consistent with the expression pattern observed in theorange leaf–Xcv interaction (Supplementary Table S2).

Microarray expression for one PR-4, RBohD and Hin-1 wasconfirmed by real-time RT-PCR analysis (Supplementary Fig. S1).Moreover, Mlo6 was represented by two different probes with sim-ilar expression ratios (Supplementary Table S2).

Genes involved in LCD

The response of orange leaves to Xcv inoculation showed HR fea-tures, suggesting the participation of LCD-related genes. Changesin gene expression were observed for five probes similar to celldeath-associated genes; three probes were down-regulated andtwo probes were up-regulated (Table 1, Supplementary Table S3).The repressed genes showed homology to two cyclic nucleotide-gated ion channels called “defense-no-death” (Genger et al., 2008)and to a red chlorophyll catabolite reductase called “acceleratedcell death 2” that catalyzes one step in the breakdown of the por-phyrin component of chlorophyll (Yao and Greenberg, 2006). Theinduced genes showed similarity to “constitutively activated celldeath 1” (Tsutsui et al., 2006) and to a cysteine protease cystatin-like inhibitor (Bozso et al., 2009). None of the five genes werepredominantly expressed in libraries obtained under stress con-ditions (Supplementary Table S3).

Alleles homologous to TFs

The citrus microarray contained 899 probes corresponding to590 Arabidopsis TFs. Of these, 58 probes analogous to 54 Arabidop-sis TFs were differentially expressed in orange leaves in response

to Xcv treatment (Supplementary Table S4). A group of 10 putativeTFs with significant over-representation in citrus stress librariesincluded one C2C2-GATA, two C2H2, two CCAAT, one HSF, two NAC(No apical meristem-ATAF-cup-shaped cotyledon) and two WRKY

d to LCD. The log2 of expression ratio between treatments (M) are shown. The M

Putative Athortholog

M ± SE

Xcv/Xcc Xcv/Ctr Xcc/Ctr

AT5G15410 −1.39 ± 0.22 −2.26 ± 0.22 −0.87 ± 0.22

AT5G54250 −1.15 ± 0.23 −0.80 ± 0.23 0.36 ± 0.23

AT4G37000 −0.99 ± 0.17 −1.12 ± 0.17 −0.13 ± 0.17

AT1G29690 1.16 ± 0.17 0.94 ± 0.17 −0.21 ± 0.17

AT3G12490 1.21 ± 0.27 1.30 ± 0.27 0.10 ± 0.27

L.D. Daurelio et al. / Journal of Plant Physiology 170 (2013) 934– 942 939

F Xcv ae etition

gttTiCCHAssp

ds(aeMr

hC(tT

D

N

emftv

ig. 5. Citrus ESTs that are differentially regulated during the non-host response toxpression ratio between treatments (M) are shown. Data are averages of three rep

enes (Fig. 5). All of these putative TFs were induced in responseo Xcv treatment, except both CCAAT genes (Fig. 5, Supplemen-ary Table S4). Phylogenetic clustering of citrus and ArabidopsisF domains showed that C2C2-GATA clustered with GATA-5-6-7nside subfamily I, both C2H2 genes clustered with ZAT6 and STZ,CAAT-HAP3 clustered with NF-YB3 within the NF-YB subfamily,CAAT-HAP5 clustered with NF-YC11 inside the NF-YC subfamily,SF clustered with HSFB1 of Group B, NAC genes clustered withNAC047 in the NAP subfamily and with ATAF1-2 in the ATAF1ubfamily, and the WRKY genes grouped with AtWRKY11-17 insideubfamily II-d and with AtWRKY42-47 inside subfamily II-b (Sup-lementary Fig. S2).

Twenty six additional TFs that belong to families related toefense responses (Singh et al., 2002) showed differential expres-ion in the citrus non-host response as follows: ten WRKY geneseight induced, one repressed and one similarly induced in Xcvnd Xcc treatments), five genes similar to the AP2-EREBP Ethyl-ne Response Factor subgroup (one repressed, four induced), fiveyb genes (four repressed, one induced) and four bZip genes (one

epressed, three induced) (Supplementary Table S4).The following genes with altered expression in microarray

ybridization analyses were confirmed by real-time RT-PCR: one2C2-GATA gene, one HSF gene, one Myb gene and two WRKY genesSupplementary Fig. S1). In addition, six genes were represented bywo different probes with similar expression ratios (Supplementaryable S4).

iscussion

on-host response to Xcv in citrus shows HR-like symptoms

The non-host response against potential pathogens is the broad-st and most important mechanism of plant defense, but it has

ainly been studied in model plants. Citrus is the most important

ruit crop in the world; however, the citrus non-host response andhe associated transcriptome have not yet been studied. In a pre-ious study, we observed that a response similar to the HR was

nd correspond to TFs with significant representation in stress libraries. The log2 ofs. Bars represent SE.

induced in orange leaves by Xcv (Daurelio et al., 2009a). In thiswork, we studied the orange response to Xcv inoculations and theresulting transcriptome modifications. Because Xcc belongs to thesame genus as Xcv, Xcc was used as the disease reference.

The orange leaves showed HR after Xcv infection. HR is a fastmacroscopic collapse of infected tissue due to LCD (Mur et al., 2008).This symptom was directly observed at high and intermediateconcentrations of bacterial inoculums but not at the lower con-centration, suggesting that the intensity of the response dependson the initial concentration of bacteria. This phenotype was differ-ent than the canker symptoms seen in Xcc infection. The chlorosisobserved at the lowest Xcv concentration and in Xcc inoculationscan be considered to be non-specific or generated by the inocu-lation method because this phenotype also appeared in Ctr. Theonset of HR-like and water soaking symptoms is fast, but this couldbe due to the young age of the leaves used for inoculation in theseexperiments. We have previously observed that symptoms appearearlier and more consistent in younger leaves compared to olderleaves. The lysis of mesophyll cells, considered a typical symptomof the HR response (Danon et al., 2000), confirmed the phenotypeobserved.

The decrease in Xcv populations observed in planta can be con-sidered typical in a non-host response (Daurelio et al., 2009b).This decrease was proportional to the initial bacterial concentra-tion, suggesting that the activation of the defense response wasmore aggressive when greater numbers of bacteria are inoculated.Because the bacterial inoculums were infiltrated into mesophylltissue, this response occurred independently of passive and pre-formed barriers. The determination of ion leakage and H2O2production, indicators of plant HR symptoms (Daurelio et al.,2009b; Rybak et al., 2009), allowed us to confirm that the citrusresponse to Xcv treatment involved an active mechanism. Thismechanism was considered to be a non-host response mediatedby HR. Ion leakage values observed during Xcv treatment were

comparable to those generated by SA used as control of cell death(Vlot et al., 2009). As seen in the grapefruit response to Xcc Aw,citrus phenotypic symptoms were slower than those observed for

9 lant P

ose

In

tohossmdgtTwa

winae

GX

mguio2t(w(fihcrc

eiootn

Ris(e(gepd

40 L.D. Daurelio et al. / Journal of P

ther plant–pathogen interactions (Rybak et al., 2009). This findinguggests the participation of a mechanism that slows the negativeffects of cell death.

dentification of genes that are expressed differentially duringon-host response in citrus

One of the greatest challenges in plant science is the identifica-ion of plant transcriptome modifications responsible for defenser disease patterns. To find genes that trigger the citrus non-ost response, transcriptome analysis was performed at 8 hpi. Webserved that oxidative burst and membrane cell disruption (earlyymptoms) were occurring at this time, while necrosis and cell lysisymptoms manifested later. The comparison of Xcv and Xcc treat-ents allowed us to detect genes specifically involved in the citrus

efense response. The higher number of differentially expressedenes in response to Xcv treatment was consistent with the quan-itative nature of plant responses to pathogens (Tao et al., 2003).he number of genes that were up-regulated and down-regulatedere similar, indicating an equivalent requirement of repression

nd induction during the defense response.The transcription profiles determined by microarray analysis

ere confirmed using real-time RT-PCR. We attributed differencesn absolute values to the dissimilar sensitivity of the two tech-iques. The validation of the microarray hybridization analysis waslso supported for those genes with two probes that present similarxpression values.

enes related to biotic stress validate that the citrus response tocv is a non-host HR

The characterization of the citrus response to the Xcv treat-ent was confirmed by the differential expression of several

enes related to the defense process. A homolog to Hin-1 wasp-regulated; Hin-1 is a cell death-related gene that is typically

nduced in the non-host HR (Oh et al., 2006). In addition, a setf PRs associated with the non-host response (van Loon et al.,006) were identified. The induced PRs were homologous to pro-eins that have defined roles in the protection against pathogensvan Loon et al., 2006). The Arabidopsis ortholog of citrus PR-15,hich belongs to a receptor family involved in the stress response

Membre et al., 2000), is also repressed during biotic stress. Thisnding suggests that the protein participates in the citrus non-ost response. Although an in silico analysis has identified PRs initrus libraries (Campos et al., 2007), this work constitutes the firsteport describing their participation in the non-host response initrus.

The Arabidopsis Mlo6 and Mlo12 genes are induced in thearly response to flagellin (Navarro et al., 2004), while Mlo12s also induced during non-host HR (Tao et al., 2003). Similarly,rthologous genes in citrus were induced. Mlo6 was found to bever-represented in citrus stress libraries, emphasizing its role inhe defense response and possible involvement in pathogen recog-ition.

The homology and expression patterns observed for PAD4,BohD and Ve2-like sequences identify them as the correspond-

ng citrus alleles. The Arabidopsis PAD4 gene is involved in SAignaling during gene-mediated and basal plant disease resistanceFeys et al., 2001) and is induced in response to flagellin (Navarrot al., 2004). RBohD is involved in extracellular ROS production

Torres et al., 2002) and triggers death in cells damaged by fun-al infection while inhibiting death in neighboring cells (Poganyt al., 2009). The Ve2 allele from Solanum lycopersicum does notarticipate in the response to pathogen attack and was represseduring an incompatible interaction (Fradin et al., 2009).

hysiology 170 (2013) 934– 942

Genes associated with LCD could explain the citrus HR phenotype

The Arabidopsis LCD genes described in detail allowed usto investigate the citrus HR phenotype. The “defense-no-death”mutants carry out defense with a greatly reduced HR and have con-stitutively elevated SA levels (Genger et al., 2008). The “acceleratedcell death 2” mutant shows the spontaneous spreading of cell deathlesions and the constitutive activation of plant defense, includingSA accumulation (Yao and Greenberg, 2006). The “constitutivelyactivated cell death 1” mutant mimics HR lesions (induction of PRgenes and increased SA concentration). This gene, induced by SA,regulates by negative feedback the SA-mediated defense pathwayof cell death (Tsutsui et al., 2006). A cystatin-like gene is inducedduring defense responses to reduce the extent of LCD (Bozso et al.,2009).

In the citrus non-host response to Xcv treatment, the down-regulation of “defense-no-death” and “accelerated cell death 2”genes should lead to an increase in the SA concentration and trig-ger LCD. The increase in SA should induce “constitutively activatedcell death 1”, which would negatively control SA and LCD levels.While down-regulation of “accelerated cell death 2′′ leads to lesionsof cell death, the diminishing of HR symptoms could be a conse-quence of “defense-no-death” down-regulation and cystatin-likeup-regulation. This observation suggests that an improved defensemechanism diminishes deleterious effects.

LCD-related genes were not significantly expressed in citrusstress libraries, indicating their participation in other physiologicalprocesses.

TFs are candidates to trigger the non-host response in citrus

A group of differentially expressed TFs that are potentiallyinvolved in the onset of the citrus non-host response were detected.The most prominent candidates were those TFs that were signifi-cantly over-represented in citrus stress libraries and that showedspecificity in the stress process. Those TFs related to plant defensemay contribute in a non-specific way.

WRKY TFs are considered to be fundamental to plant defense(Eulgem, 2005). WRKY genes were the largest group of TFs differ-entially expressed during the citrus non-host response. Two WRKYTFs were induced and over-represented in citrus stress libraries.The aCL4Contig6 allele was similar to AtWRKY11-17, which arenegative regulators of the basal response and inductors of genesrelated to jasmonic acid (JA) defense (Journot-Catalino et al., 2006).The aCL1201Contig1-2 allele was similar to the uncharacterizedAtWRKY31 gene but grouped inside the subfamily IIb in whichAtWRKY72 has been implicated in disease resistance (Bhattaraiet al., 2010).

The aCL1506Contig1-aIC0AAA2BH06RM1 c allele was similarto AtWRKY33, a pathogen-induced TF required for resistance tonecrotrophic fungal pathogens (Zheng et al., 2006). Three alleles(aCL775Contig1, aCL3000Contig1, and aIC0AAA56BD05RM1 c)were similar to the pathogen-induced gene AtWRKY53, whichactivates the defense response (Wang et al., 2006). These threegenes were grouped in subfamily III with the pathogen-inducedAtWRKY54-70, defined as negative regulators of SA synthesis(Wang et al., 2006). The aCL2994Contig1 allele was clustered insubfamily II-a with the pathogen-induced genes AtWRKY18-40-60, which have been shown to have redundant negative effects onSA and positive roles in JA-mediated defense (Xu et al., 2006). TheaCL775Contig1 allele had similar expression levels in both bacte-rial treatments and could be involved in the basal response. This

allele contains a WRKY domain similar to AtWRKY33, which regu-lates the antagonistic relationship between the defense responsesto Pseudomonas syringae and necrotrophic fungal pathogens. TheaC05807G12SK c allele was similar to the AtWRKY75 and grouped

lant P

irbdt

rdadtaeTrwpt

dcatiAwtacbT(srnhc

AteOstetv

rmtttrsdrh

A

PoE

L.D. Daurelio et al. / Journal of P

nside subfamily IIb, which has not been described yet. The onlyepressed WRKY allele (aCL1204Contig3) was similar to AtWRKY20ut grouped in subfamily I near AtWRKY58, a negative regulator ofisease resistance (Wang et al., 2006). This finding coincides withhe expression pattern observed.

The citrus Myb genes detected in the transcriptome wereelated to Arabidopsis orthologs that are not associated with theefense response. The Myb domain of the induced aC05146C04SK cllele was homologous to NtMyb2, which positively regulatesefense genes and is itself regulated during wounding and elicitorreatment (Sugimoto et al., 2003). NtMyb2 was transcriptionallyctivated by the GATA-type zinc finger protein AGP1 (Sugimotot al., 2003). The AGP1 gene is homologous to the aCL908Contig1F, which was induced in response to Xcv treatment and was over-epresented in citrus stress libraries. AGP1 induction by woundingas previously reported in tobacco (Sugimoto et al., 2003), but itsarticipation in the non-host response to bacteria is reported forhe first time here.

In addition to the WRKY and GATA TFs mentioned above, sevenifferentially expressed citrus TFs showed over-representation initrus stress libraries. The induced C2H2 alleles (aCL382Contig2 andCL146Contig3) were similar to the Arabidopsis STZ Zn finger pro-ein that controls JA-regulated genes (Pauwels et al., 2008). Thenduced NAC allele aCL157Contig4-aC19004C08T7 c was similar totATAF1; this citrus allele grouped within the same subfamily butas nearest to the ATAF2, which is associated with regulation of

he defense response in pepper (Oh et al., 2005). The induced NACllele aCL267Contig2-4 has homology to a previously characterizeditrus TF that was induced in different stresses (Fan et al., 2007)ut not previously identified in the non-host defense response.he Arabidopsis alleles related to HSF (aCL432Contig1) and CCAATaCL1964Contig1 and aKN0AAQ5YA18RM1 c) were not previouslytudied during the non-host response. CCAAT genes were over-epresented in stress libraries but were repressed during the citruson-host response. This observation indicates that these genesave a possible ambiguous or specific function and need furtherharacterization.

The induced AP2-EREBP allele aCL8155Contig1 was similar totERF1, which is induced during the Arabidopsis defense response

o bacteria and has been proposed to be a connection between thethylene and JA pathways in plant defense (Lorenzo et al., 2003).n the other hand, the induced AP2-EREBP aCL152Contig1 was

imilar to AtERF72. AtERF72 is induced in response to infection byhe incompatible fungal pathogen Alternaria brassicicola (Okamurot al., 1997). The aC08022H08SK c TF showed sequence similarity tohe bZIP gene At1HY5, a negative regulator of ethylene biosynthesisia the activation of AtERF11.

In conclusion, we have characterized the citrus non-hostesponse to Xcv treatment and obtained evidence that the responseechanism is similar to HR. Transcriptional profiling supported

his hypothesis and further explained the phenotype observed. Theranscriptional survey and the in silico analysis of citrus EST dis-ribution allowed us to identify several TFs that are potentiallyesponsible for triggering the non-host response in citrus. Theseelected TFs could be used in molecular breeding to improve plantefense. We identified potential regulatory genes in the non-hostesponse in citrus that are novel. These data make it possible toypothesize about this mechanism in other woody crop species.

cknowledgements

This work was supported by grants from the Agencia Nacional deromoción Científica y Tecnológica (ANPCyT BID-PICT-2010-1762)f Argentina and from the Programa de Cooperación Argentina-spana (ES/09/10, Ministerio de Ciencia, Tecnología e Innovación

hysiology 170 (2013) 934– 942 941

Productiva of Argentina, Ministerio de Ciencia e Innovación ofSpain) to EGO and from the Subprograma de Acciones Integradas- Ministerio de Ciencia e Innovación (AR2009-0023) of Espana toFRT. EGO and LDD are staff members and SP is a fellow of the Con-sejo Nacional de Investigaciones Científicas y Técnicas (CONICET,Argentina).

The authors thank the International Union of Biochemistry andMolecular Biology Wood-Whelan fellowship program for fundingLDD’s stay at IVIA. We are also grateful to Professors Mercedes Leivaand Hebe Bottai for their assistance with statistical analysis, to thetechnician Sebastián Graziati for support with plant materials andto Dr. Daniel G. Kurth for his help with Perl programming.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.jplph.2013.01.011.

References

Agusti J, Merelo P, Cercos M, Tadeo FR, Talon M. Ethylene-induced differential geneexpression during abscission of citrus leaves. J Exp Bot 2008;59:2717–33.

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, et al. Gapped BLASTand PSI-BLAST: a new generation of protein database search programs. NucleicAcids Res 1997;25:3389–402.

Ascencio-Ibanez JT, Sozzani R, Lee TJ, Chu TM, Wolfinger RD, Cella R, et al. Globalanalysis of Arabidopsis gene expression uncovers a complex array of changesimpacting pathogen response and cell cycle during geminivirus infection. PlantPhysiol 2008;148:436–54.

Bhattarai KK, Atamian HS, Kaloshian I, Eulgem T. WRKY72-type transcription factorscontribute to basal immunity in tomato and Arabidopsis as well as gene-for-generesistance mediated by the tomato R gene Mi-1. Plant J 2010;63:229–40.

Bozso Z, Maunoury N, Szatmari A, Mergaert P, Ott PG, Zsiros LR, et al. Transcriptomeanalysis of a bacterially induced basal and hypersensitive response of Medicagotruncatula. Plant Mol Biol 2009;70:627–46.

Campos MA, Rosa DD, Texeira JEC, Targon ML, Souza AA, Stach-Machado DR, et al. PRgene families of citrus: their organ specific-biotic and abiotic inducible expres-sion profiles based on ESTs approach. Genet Mol Biol 2007;30:917–30.

Cercós M, Soler G, Iglesias DJ, Gadea J, Forment J, Talon M. Global analysis of geneexpression during development and ripening of citrus fruit flesh. A proposedmechanism for citric acid utilization. Plant Mol Biol 2006;62:513–27.

Cernadas RA, Camillo LR, Benedetti CE. Transcriptional analysis of the sweet orangeinteraction with the citrus canker pathogens Xanthomonas axonopodis pv. citriand Xanthomonas axonopodis pv. aurantifolii. Mol Plant Pathol 2008;9:609–31.

D’Ambrogio de Argüeso A. Manual de técnicas en histología vegetal. Buenos Aires:Editorial Hemisferio Sur S.A. 1986.

Danon A, Delorme V, Mailhac N, Gallois P. Plant programmed cell death: a commonway to die. Plant Physiol Biochem 2000;38:647–55.

Daurelio LD, Checa SK, Barrio JM, Ottado J, Orellano EG. Characterization of Cit-rus sinensis type 1 mitochondrial alternative oxidase and expression analysisin biotic stress. Biosci Rep 2009a;30:59–71.

Daurelio LD, Tondo ML, Dunger G, Gottig N, Ottado J, Orellano EG. Hypersensitiveresponse. In: Narwal SS, Catalán AN, Sampietro DA, Vattuone MA, Polyticka B,editors. Book on Plant Bioassays. Houston: Studium Press; 2009b. p. 187–206.

Daurelio LD, Petrocelli S, Blanco F, Holuigue L, Ottado J, Orellano EG. Transcriptomeanalysis reveals novel genes involved in nonhost response to bacterial infectionin tobacco. J Plant Physiol 2011;168:382–91.

Eulgem T. Regulation of the Arabidopsis defense transcriptome. Trends Plant Sci2005;10:71–8.

Fan J, Gao X, Yang YW, Deng W, Li ZG. Molecular cloning and characterization ofa NAC-like gene in “Navel” orange fruit response to postharvest stresses. PlantMol Biol Rep 2007;25:145–53.

Feys BJ, Moisan LJ, Newman MA, Parker JE. Direct interaction between the Arabidopsisdisease resistance signaling proteins, EDS1 and PAD4. EMBO J 2001;20:5400–11.

Finn RD, Mistry J, Schuster-Bockler B, Griffiths-Jones S, Hollich V, Lassmann T, et al.Pfam: clans, web tools and services. Nucleic Acids Res 2006;34:D247–51.

Forment J, Gadea J, Huerta L, Abizanda L, Agusti J, Alamar S, et al. Developmentof a citrus genome-wide EST collection and cDNA microarray as resources forgenomic studies. Plant Mol Biol 2005;57:375–91.

Fradin EF, Zhang Z, Juarez Ayala JC, Castroverde CD, Nazar RN, Robb J, et al. Geneticdissection of Verticillium wilt resistance mediated by tomato Ve1. Plant Physiol2009;150:320–32.

Fujii H, Shimada T, Sugiyama A, Nishikawa F, Endo T, Nakano M, et al. Profiling

ethylene-responsive genes in mature mandarin fruit using a citrus 22K oligoar-ray. Plant Sci 2007;173:340–8.

Gandia M, Conesa A, Ancillo G, Gadea J, Forment J, Pallas V, et al. Transcriptio-nal response of Citrus aurantifolia to infection by Citrus tristeza virus. Virology2007;367:298–306.

9 lant P

G

G

G

G

H

J

J

K

L

M

M

M

M

N

N

O

O

O

P

P

42 L.D. Daurelio et al. / Journal of P

enger RK, Jurkowski GI, McDowell JM, Lu H, Jung HW, Greenberg JT, et al. Signalingpathways that regulate the enhanced disease resistance of Arabidopsis “defense,no death” mutants. Mol Plant Microbe Interact 2008;21:1285–96.

entleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, et al. Bioconduc-tor: open software development for computational biology and bioinformatics.Genome Biol 2004;5:R80.

omi K, Yamamato H, Akimitsu K. Epoxide hydrolase: a mRNA induced by the fungalpathogen Alternaria alternata on rough lemon (Citrus jambhiri). Plant Mol Biol2003;53:189–99.

uidetti-Gonzalez S, Freitas-Astúa J, Morais do Amaral A, Martins NF, Mehta A, San-tos Silva M, et al. Genes associated with hypersensitive response (HR) in thecitrus EST database (CitEST). Genet Mol Biol 2007;30:943–56.

uang X, Madan A. CAP3: a DNA sequence assembly program. Genome Res1999;9:868–77.

ohansen DA. Plant Microtechnique 1662. New York and London: McGraw-Hill BookCo; 1940.

ournot-Catalino NI, Somssich IE, Roby D, Kroj T. The transcription factors WRKY11and WRKY17 act as negative regulators of basal resistance in Arabidopsisthaliana. Plant Cell 2006;18:3289–302.

im JS, Sagaram US, Burns JK, Li JL, Wang N. Response of sweet orange (Citrus sinen-sis) to ‘Candidatus Liberibacter asiaticus’ infection: microscopy and microarrayanalyses. Phytopathology 2009;99:50–7.

orenzo O, Piqueras R, Sanchez-Serrano JJ, Solano R. Ethylene Response Factor1 inte-grates signals from ethylene and jasmonate pathways in plant defense. Plant Cell2003;15:165–78.

artinez-Godoy MA, Mauri N, Juarez J, Marques MC, Santiago J, Forment J, et al.A genome-wide 20K citrus microarray for gene expression analysis. BMCGenomics 2008;9:318.

embre N, Bernier F, Staiger D, Berna A. Arabidopsis thaliana germin-like pro-teins: common and specific features point to a variety of functions. Planta2000;211:345–54.

ozoruk J, Hunnicutt LE, Cave RD, Hunter WB, Bausher MG. Profiling transcrip-tional changes in Citrus sinensis (L.) Osbeck challenged by herbivory from thexylem-feeding leafhopper Homalodisca coagulata (Say) by cDNA macroarrayanalysis. Plant Sci 2006;170:1068–80.

ur LA, Kenton P, Lloyd AJ, Ougham H, Prats E. The hypersensitive response; thecentenary is upon us but how much do we know? J Exp Bot 2008;59:501–20.

avarro L, Zipfel C, Rowland O, Keller I, Robatzek S, Boller T, et al. The trans-criptional innate immune response to flg22. Interplay and overlap with Avrgene-dependent defense responses and bacterial pathogenesis. Plant Physiol2004;135:1113–28.

ürnberger T, Lipka V. Non-host resistance in plants: new insights into an old phe-nomenon. Mol Plant Pathol 2005;6:335–45.

h SK, Lee S, Yu S, Choi D. Expression of a novel NAC domain-containing transcrip-tion factor (CaNAC1) is preferentially associated with incompatible interactionsbetween chili pepper and pathogens. Planta 2005;222:876–87.

h SK, Lee S, Chung E, Park JM, Yu SH, Ryu CM, et al. Insight into types I and II nonhostresistance using expression patterns of defense-related genes in tobacco. Planta2006;223:1101–7.

kamuro JK, Caster B, Villarroel R, Van MM, Jofuku KD. The AP2 domain of APETALA2defines a large new family of DNA binding proteins in Arabidopsis. Proc Natl AcadSci USA 1997;94:7076–81.

auwels L, Morreel K, De Witte E, Lammertyn F, Van Montagu M, Boerjan W,et al. Mapping methyl jasmonate-mediated transcriptional reprogramming of

metabolism and cell cycle progression in cultured Arabidopsis cells. Proc NatlAcad Sci USA 2008;105:1380–5.

ogany M, von RU, Grun S, Dongo A, Pintye A, Simoneau P, et al. Dual roles of reac-tive oxygen species and NADPH oxidase RBOHD in an Arabidopsis–Alternariapathosystem. Plant Physiol 2009;151:1459–75.

hysiology 170 (2013) 934– 942

Romualdi C, Bortoluzzi S, D’Alessi F, Danieli GA. IDEG6: a web tool for detectionof differentially expressed genes in multiple tag sampling experiments. PhysiolGenomics 2003;12:159–62.

Rozen S, Skaletsky H. Primer3 on the WWW for general users and for biologistprogrammers. In: Krawetz S, Misener S, editors. Bioinformatics methods andprotocols: methods in molecular biology. Totowa: Humana Press; 2000. p.365–86.

Rybak M, Minsavage GV, Stall RE, Jones JB. Identification of Xanthomonas citri ssp.citri host specificity genes in a heterologous expression host. Mol Plant Pathol2009;10:249–62.

Saitou N, Nei M. The neighbor-joining method: a new method for reconstructingphylogenetic trees. Mol Biol Evol 1987;4:406–25.

Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: a laboratory manual. NewYork: Cold Spring Harbor Laboratory Press; 1989.

Singh K, Foley RC, Onate-Sanchez L. Transcription factors in plant defense and stressresponses. Curr Opin Plant Biol 2002;5:430–6.

Smith GK. Linear models and empirical Bayes methods for assessing differen-tial expression in microarray experiments. Stat Appl Genet Mol Biol 2004;3,Article 3.

Stothard P. The sequence manipulation suite: JavaScript programs for analyzing andformatting protein and DNA sequences. Biotechniques 2000;28, 1102, 04.

Sugimoto K, Takeda S, Hirochika H. Transcriptional activation mediated by bindingof a plant GATA-type zinc finger protein AGP1 to the AG-motif (AGATCCAA) ofthe wound-inducible Myb gene NtMyb2. Plant J 2003;36:550–64.

Talon M, Gmitter Jr FG. Citrus genomics. Int J Plant Sci 2008;2008:528361.Tamura K, Dudley J, Nei M, Kumar S. MEGA4: Molecular Evolutionary Genetics Anal-

ysis (MEGA) software version 4.0. Mol Biol Evol 2007;24:1596–9.Tao Y, Xie Z, Chen W, Glazebrook J, Chang HS, Han B, et al. Quantitative nature of

Arabidopsis responses during compatible and incompatible interactions with thebacterial pathogen Pseudomonas syringae. Plant Cell 2003;15:317–30.

Thieme F, Koebnik R, Bekel T, Berger C, Boch J, Buttner D, et al. Insights into genomeplasticity and pathogenicity of the plant pathogenic bacterium Xanthomonascampestris pv. vesicatoria revealed by the complete genome sequence. J Bacteriol2005;187:7254–66.

Thimm O, Blasing O, Gibon Y, Nagel A, Meyer S, Kruger P, et al. MAPMAN: auser-driven tool to display genomics data sets onto diagrams of metabolic path-ways and other biological processes. Plant J 2004;37:914–39.

Torres MA, Dangl JL, Jones JD. Arabidopsis gp91phox homologues AtrbohD and Atr-bohF are required for accumulation of reactive oxygen intermediates in the plantdefense response. Proc Natl Acad Sci USA 2002;99:517–22.

Tsutsui T, Morita-Yamamuro C, Asada Y, Minami E, Shibuya N, Ikeda A, et al. Sali-cylic acid and a chitin elicitor both control expression of the CAD1 gene involvedin the plant immunity of Arabidopsis. Biosci Biotechnol Biochem 2006;70:2042–8.

van Loon LC, Rep M, Pieterse CM. Significance of inducible defense-related proteinsin infected plants. Annu Rev Phytopathol 2006;44:135–62.

Vlot AC, Dempsey DA, Klessig DF. Salicylic acid: a multifaceted hormone to combatdisease. Annu Rev Phytopathol 2009;47:177–206.

Wang D, Amornsiripanitch N, Dong X. A genomic approach to identify regulatorynodes in the transcriptional network of systemic acquired resistance in plants.PLoS Pathog 2006;2:e123.

Xu X, Chen C, Fan B, Chen Z. Physical and functional interactions betweenpathogen-induced Arabidopsis WRKY18, WRKY40, and WRKY60 transcriptionfactors. Plant Cell 2006;18:1310–26.

Yao N, Greenberg JT. Arabidopsis accelerated cell death 2 modulates programmedcell death. Plant Cell 2006;18:397–411.

Zheng Z, Qamar SA, Chen Z, Mengiste T. Arabidopsis WRKY33 transcriptionfactor is required for resistance to necrotrophic fungal pathogens. Plant J2006;48:592–605.